Active psychosis in schizophrenia is among the most severe and burdensome medical conditions worldwide. However, the mechanisms of psychotic symptoms in this disorder, such as hallucinations (i.e., abnormal percepts in the absence of external sensory stimuli), remain elusive. This K23 application presents a research and training program that will support the applicant on a path towards becoming an NIH-funded independent investigator focused on the application of functional neuroimaging to the study of psychotic symptoms in schizophrenia. The activities in this application build on the candidate's prior training and are set in a resource-rich environment that will foster his development of expertise in (1) advanced analytic methods and study conduct for functional magnetic resonance imaging (fMRI) research; (2) computational neuroscience; (3) perception and cognition research; (4) pathophysiology and clinical assessment of schizophrenia; and (5) responsible and ethical conduct in scientific research with vulnerable populations. Combining functional magnetic resonance imaging (fMRI) and computational modeling, the current research proposal seeks to (1) define the neural mechanisms that generate hallucinations in schizophrenia; and (2) inform the development of a computational model of hallucinations based on predictive coding, an empirically-validated theoretical framework that supports a role of sensory systems in learning and predicting regularities in the external environment. The overarching hypothesis is that abnormal prediction-based attenuation of sensory cortical function produces excessive activity in the sensory cortex that generates hallucinations. To test this hypothesis, the present study will employ a novel speech discrimination fMRI paradigm, two groups of patients with schizophrenia, those with active, frequent auditory verbal hallucinations and those without a significant history of hallucinations, and a third group of healthy controls. This design will allo for testing a direct link between dysfunction in sensory predictive-coding mechanisms and the online experience of hallucinations in patients with schizophrenia, and will thus inform the neurobiological basis of psychotic symptoms in this disorder. Together, this training and research program will facilitate the candidate's transition to an independent research career and will help identify new therapeutic targets for refractory psychosis. RELEVANCE: The novel application of the predictive-coding framework and model-based fMRI to the study of psychotic symptoms will shed new light on the mechanisms of generation of psychotic symptoms, thus filling an important gap in schizophrenia research. This project will serve to develop an explanatory model of hallucinations that can be used to generate specific, testable hypotheses for future neuroscience research in both humans and non-human animal models, and to uncover novel targets (sensory prediction deficits) likely modifiable by treatment via learning or pharmacotherapy.